Vertical Structure of an OGCM Simulation of the Equatorial Pacific Ocean in 1985–94

Abstract

The vertical structure of the variability in the equatorial Pacific in a high-resolution ocean general circulation model (OGCM) simulation for 1985?94 is investigated. Near the equator the linear vertical modes are estimated at each grid point and time step of the OGCM simulation. The characteristics of the vertical modes are found to vary more in space than in time. The contribution of baroclinic modes to surface zonal current and sea level anomalies is analyzed. The first two modes contribute with comparable amplitude but with different spatial distribution in the equatorial waveguide. The third and fourth modes exhibit peaks in variability in the east and in the westernmost part of the basin where the largest zonal gradients in the density field and in the vertical mode characteristics are found. Higher-order mode (sum of third to the eighth mode) variability is the largest near the date line close to the maximum in zonal wind stress variability. Kelvin and first-meridional Rossby components are derived for each of the first three baroclinic contributions by projection onto the associated meridional structures. They are compared to equivalent ones in multimode linear simulations done with the projection coefficients and phase speeds derived from the OGCM simulation. This suggests that in addition to the first-mode-forced equatorial Kelvin and Rossby waves earlier found in the data, forced waves of higher vertical modes should also be observable. For the first two vertical modes, the anomalies in the linear and the OGCM simulations have a similar magnitude and usually present similar propagation characteristics. Phase speed characteristics are however different in the eastern Pacific with larger values for the OGCM. The effect of zonal changes in the stratification is tested in the linear model for a stratification change located either in the eastern or in the western Pacific. This results in a significant redistribution of energy to higher modes via modal dispersion. In particular the third mode increases to a magnitude closer to the one in the OGCM simulation. Gravest modes are also affected. This suggests that modal dispersion plays an important role and should be considered when interpreting data as a combination of linear long equatorial waves.